How do bike-sharing usage patterns evolve before and after the pandemic? Evidence from the city of chicago

The COVID-19 pandemic significantly altered travel behaviors, leading to notable evolutions in bike-sharing usage patterns. With limited explorations of post-pandemic patterns in existing literature, this study addresses two questions: (1) What are the spatiotemporal features of bike-sharing ridership before and after the COVID-19 pandemic? (2) What factors influenced bike-sharing travel behaviors during these periods? To answer the first question, we analyze bike-sharing trip data in Chicago from 2018 to 2023 (spanning pre-, during, and post-pandemic), to provide descriptive analysis of the overall, temporal, and spatial shifts in bike-sharing usage. For the second question, by integrating bike-sharing trip data with socio-demographics, land use, bike-sharing docks, and COVID-19 cases at the block group level, we employ Gradient Boosting Decision Tree (GBDT) models combined with SHapley Additive exPlanations (SHAP) to quantify the importance of these factors, along with their nuanced effects across samples. Findings reveal that, firstly, post-pandemic bike-sharing patterns grew on weekday evenings and weekend afternoons, and bike-sharing docks and bike-sharing usage were expanded to the surrounding neighborhoods. Secondly, White/Asian proportion, labor force, income and commercial land at the census block group level are critical influencing factors on bike-sharing usage, among which the number of docks is the most influential one. Notably, thresholds for the key factors are observed across the three periods, below which the positive and negative effects on bike-sharing usage vary among samples. Once exceeding the thresholds their impacts are commonly positive, which indicates the scale effects of these factors. The positive impacts of these factors on bike-sharing usage increased post-pandemic compared to those pre- and during-pandemic. These findings together demonstrate increased commercial and recreational use of bike-sharing and identify frequent bike-sharing user groups including the White and Asian populations in Chicago, which provide empirical evidence for bike-sharing facility planning and service operations.